NMR-spectroscopic and solid-state investigations of cometal-free asymmetric conjugate addition: A dinuclear paracyclophaneimine zinc methyl complex

Article abstract of DOI:10.1021/ja1032502

We present herein the first indications for dimeric structures in cometal-free asymmetric conjugate addition reactions of dialkylzinc reagents with aldehydes. These are revealed by nonlinear effect (NLE) studies. A monomer-dimer equilibrium can be assumed which explains the increase of the ee value in the product over time. Also, DOSY NMR spectroscopic measurements indicate the existence of the catalyst as [LZnEt]_n complexes in solution. Additionally, the first X-ray structure of a zinc complex with a [2.2]paracyclophane ligand was determined. The structures of the zinc complexes are supported by DFT calculations of monomeric and dimeric species.

Full text of DOI:10.1021/ja1032502

Published on Web 08/31/2010
NMR-Spectroscopic and Solid-State Investigations of
Cometal-Free Asymmetric Conjugate Addition: A Dinuclear
Paracyclophaneimine Zinc Methyl Complex
S. Ay,† R. E. Ziegert,† H. Zhang,‡ M. Nieger,§ K. Rissanen,0 K. Fink,¶ A. Kubas,¶
R. M. Gschwind,*,‡ and S. Bra¨se*,†
Institute of Organic Chemistry, Karlsruhe Institute of Technology (KIT), Fritz-Haber-Weg 6,
D-76131 Karlsruhe, Germany, Institute of Organic Chemistry, UniVersity of Regensburg,
D-93040 Regensburg, Germany; Laboratory of Inorganic Chemistry, UniVersity of Helsinki,
P.O. Box 55 (A.I. Virtasen aukio 1), FIN-00014 UniVersity of Helsinki, Finland; Laboratory of
Organic Chemistry, UniVersity of JyVa¨skyla¨, P.O. Box 35, FIN-40014 UniVersity of JyVa¨skyla¨,
Finland; Institute of Nanotechnology (INT), Karlsruhe Institute of Technology (KIT),
Hermann-Von-Helmholtz-Platz 1, D-76344 Eggenstein-Leopoldshafen, Germany
Received April 23, 2010; E-mail: ruth.gschwind@chemie.uni-regensburg.de; braese@kit.edu
Abstract: We present herein the first indications for dimeric structures in cometal-free asymmetric conjugate
addition reactions of dialkylzinc reagents with aldehydes. These are revealed by nonlinear effect (NLE)
studies. A monomer-dimer equilibrium can be assumed which explains the increase of the ee value in the
product over time. Also, DOSY NMR spectroscopic measurements indicate the existence of the catalyst
as [LZnEt]n complexes in solution. Additionally, the first X-ray structure of a zinc complex with a
[2.2]paracyclophane ligand was determined. The structures of the zinc complexes are supported by DFT
calculations of monomeric and dimeric species.
1. Introduction
toward a wide range of functional groups. Typically, organo-zinc
compounds react with electrophiles in 1,2-addition reactions.5
The postulated mechanism is investigated experimentally and
theoretically.6 In contrast, no mechanistic investigations have
been made so far to explain the unexpected 1,4-selectivity of
our [2.2]paracyclophane-based system reacting dialkylzinc
reagents with enals.
Herein we present the first mechanistic insights in this
direction. Besides nonlinear effect (NLE) studies, zinc catalyst
intermediates of the cometal-free ACA were studied by NMR
techniques in solution. Furthermore, we present the first X-ray
structure of a zinc-[2.2]paracyclophane complex. Finally, DFT
calculations supported our findings.
The asymmetric conjugate addition (ACA) reaction is one
of the most powerful C-C bond-forming reactions in modern
organic chemistry.1 Beside existing copper- or rhodium-
catalyzed additions of dialkylzinc reagents, our group showed
that [2.2]paracyclophane-based N,O-ligands allow an efficient
zinc-catalyzed alternative, which occurs without additional metal
salts (cometals are copper or rhodium in addition to the
organometallic reagents such as dialkylzinc or Grignard re-
agents).2 Moreover, this system is compatible with enals,
whichsdue to the high electrophilicity of the aldehyde
groupsoften cause chemoselectivity problems with other cata-
lytic systems. Although efficient systems were developed to react
Grignard3 and aluminum4 reagents in 1,4-addition reactions, zinc
reagents are still very attractive due to their high tolerance
2. Results and Discussion
Recently, we were able to optimize the [2.2]paracyclophane
ligand structure for conjugate addition reactions of diethylzinc
† Organic Chemistry, KIT.
‡ University of Regensburg.
§ Laboratory of Inorganic Chemistry, University of Helsinki.
0 Laboratory of Organic Chemistry, University of Jyva¨skyla¨.
¶ INT, KIT.
(1) (a) Lo´pez, F.; Feringa, B. L. In Asymmetric Synthesis: The Essentials;
Christmann, M., Bra¨se, S., Eds.; Wiley-VCH: Weinheim, 2008, pp
78-83. (b) Alexakis, A.; Benhaim, C. Eur. J. Org. Chem. 2002, 3221–
3236. (c) Krause, N.; Hoffmann-Ro¨der, A. Synthesis 2001, 171–196.
(d) Perlmutter, P. Conjugate Addition Reactions in Organic Synthesis,
Tetrahedron Organic Chemistry Series, Vol. 9; Pergamon: Oxford,
1992.
(2) Bra¨se, S.; Ho¨fener, S. Angew. Chem. 2005, 117, 8091–8093. Angew.
Chem., Int. Ed. 2005, 44, 4879-7881.
(3) (a) Bos, P. H.; Minnaard, A. J.; Feringa, B. L. Org. Lett. 2008, 10,
4219–4222. (b) Harutyunyan, S. R.; den Hartog, T.; Geurts, K.;
Minnaard, A. J.; Feringa, B. L. Chem. ReV. 2008, 108, 2824–2852.
(c) Lo´pez, F.; Minnaard, A.; Feringa, B. L. Acc. Chem. Res. 2007,
40, 179–188.
(4) (a) Hawner, C.; Li, K.; Cirriez, V.; Alexakis, A. Angew. Chem. 2008,
120, 8334–8337. Angew. Chem., Int. Ed. 2008, 47, 8211-8214. (b)
d’Augustin, M.; Alexakis, A. Tetrahedron Lett. 2007, 48, 7408–7412.
(c) Alexakis, A.; Albrow, V.; Biswas, K.; d’Augustin, M.; Prieto, O.;
Woodward, S. Chem. Commun. 2005, 22, 2843–2845.
(5) Review about substrates with carbonyl moieties: Pu, L.; Yu, H.-B.
Chem. ReV. 2001, 101, 757–824. High enantioselectivities have been
obtained with N,N-(dimethylamino)isoborneol (DAIB) as ligand and
aldehydes as substrates Kitamura, M.; Okada, S.; Suga, S.; Noyori,
R. J. Am. Chem. Soc. 1989, 111, 4028–4036.
(6) (a) Yamakawa, M.; Noyori, R. J. Am. Chem. Soc. 1995, 117, 6327–
6335. (b) Yamakawa, M.; Noyori, R. Organometallics 1999, 18, 128–
133. (c) Goldfuss, B.; Houk, K. N. J. Org. Chem. 1998, 63, 8998–
9006. (d) Goldfuss, B.; Steigelmann, M.; Khan, S. I.; Houk, K. N. J.
Org. Chem. 2000, 65, 77–82. (e) Rasmussen, T.; Norrby, P.-O. J. Am.
Chem. Soc. 2001, 123, 2464–2465.
9
10.1021/ja1032502  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 12899–12905 12899

A R T I C L E S
Ay et al.
Scheme 1. ACA of Diethylzinc with Cinnamaldehyde Utilizing [2.2]Paracyclophane Ligandsa
a brsm ) based on recovered starting material.
graphic separation of the corresponding diastereomers.10 If one
diastereomer generates a catalyst which is significantly more
active, a tedious chromatographic purification step might be
avoided.11 Ligand (Rp,S)-4 was mixed with its diastereomer
(Sp,S)-4 in different de’s. The results of the NLE study are shown
in Figure 3. As for the different mixtures of enantiomers (Figure
1), a clear negatiVe NLE is observed. The experimental values
correlate again very well with a calculated fitted curve (see SI),
see also Figure 2 (LR and LS are diastereomers). Both diaster-
eomers deliver compound 2 in >99% ee. Ligand (Rp,S)-4 leads
to (S)-2, whereas (Sp,S)-4 gives (R)-2. In the case of an
equimolar mixture of both diastereomers, 2 is obtained in -36%
ee due to a higher catalytic activity of the generated zinc catalyst
with the (Sp,S)-4 ligand. The slight difference between the
experimental values and the fitted curve in comparison to Figure
1 indicates complex interactions between the diastereomeric
ligands (Vide infra).
Considering the observed NLEsbesides the different activi-
ties of diastereomeric complexes of (Rp,S)-4 and (Sp,S)-4sthe
potential formation of dimers has to be taken into account. As
the NLE is negative, homodimers are likely to be more stable
than the corresponding heterodimers (see section with calculated
values).9 Similar effects have been observed by Noyori using
diastereomeric DAIB ligands.12 In our case, however, the
specific curve progression indicates a more complex aggregation
level, which might be caused by contributions of [LZnEt]n
oligomers as indicated by NMR measurements (see below).13
Taking all considerations established so far into account, a
(ML)2-system is likely to exist.
Figure 1. Observed and predicted negative NLE for the 1,4-addition
reaction using (Rp,S)-4/(Sp,R)-4 (0.01 M). The calculated curve is based on
parameters found by fitting of the experimental values (eemax ) 99%): see
SI.
to cinnamaldehyde.7 Herein, the corresponding ligand 4 was
used for NLE-investigations,8 which proved to be a valuable
tool for mechanistic investigations. The ligand (Rp,S)-4 was used
in enantiopurities ranging from 0% ee to >99% ee with the
reaction conditions given in Scheme 1. As depicted in Figure
1, a significant negatiVe NLE was observed. For the time
dependency of the ee, see Supporting Information (SI).
The experimental values correlate very well with a fitted curve
calculated by assuming stable homo- or heterodimers as
indicated in Figure 2 (LR and LS are enantiomers; for math-
ematical models used, see SI). This is an indication for a favored
formation of dimeric species. Interestingly, ligand (Rp,S)-4
delivers also a negative NLE in 1,2-addition reactions of
dialkylzinc reagents to aldehydes.9
In order to study the catalytically active species in more detail,
NMR spectroscopic measurements were carried out. For these
NMR studies of the zinc complex structure in solution, the
phenyl-substituted ligand 5 was chosen (Scheme 1). This ligand
allows observing reliably the formation of ethylzinc complexes
1
due to fewer signals in the aliphatic region of the H NMR
In the course of our NLE studies, we also compared
diastereomers (Rp,S)-4 and (Sp,S)-4. The initially used enan-
tiopure [2.2]paracyclophane ligands are obtained by chromato-
spectrum.
(10) Bra¨se, S.; Dahmen, S.; Ho¨fener, S.; Lauterwasser, F.; Kreis, M.;
Ziegert, R. E. Synlett 2004, 2647–2669.
(7) Ay, S.; Nieger, M.; Bra¨se, S. Chem.sEur. J. 2008, 14, 11539–11556.
(8) NLE ) non linear effect: (a) Girard, C.; Kagan, H. B. Angew. Chem.
1998, 110, 3088–3127. Angew. Chem., Int. Ed. 1998, 37, 2923-2959.
(b) Kagan, H. B. AdV. Synth. Catal. 2001, 343, 227–233. (c)
Satyanarayana, T.; Abraham, S.; Kagan, H. B. Angew. Chem. 2009,
121, 464–503. Angew. Chem., Int. Ed. 2009, 48, 456-494. (d) For a
recent example of a (-)-NLE showing that the homodimer is more
stable than the heterodimer: Kina, A.; Iwamura, H.; Hayashi, T. J. Am.
Chem. Soc. 2006, 128, 3904–3905.
(11) Bolm, C.; Muniz, K.; Hildebrand, J. P. Org. Lett. 1999, 1, 491–493.
(12) Kitamura, M.; Suga, S.; Niwa, M.; Noyori, R. J. Am. Chem. Soc. 1995,
117, 4832–4842.
(13) Similar curves (using enantiomeric ligands) were obtained in Ti-
catalyzed asymmetric oxidations of sulfides (Kagan13a) and in Cu-
catalyzed ACA of Grignard reagents to enones (Pfaltz and Zhou13b).
But also in these cases the exact structure of the participating species
remained unknown due to complex equilibria: (a) Guillaneux, D.;
Zhao, S.-H.; Samuel, O.; Rainford, D.; Kagan, H. B. J. Am. Chem.
Soc. 1994, 116, 9430–9439. (b) Zhou, Q.-L.; Pfaltz, A. Tetrahedron
1994, 50, 4467–4478.
(9) Lauterwasser, F.; Vanderheiden, S.; Bra¨se, S. AdV. Synth. Catal. 2006,
348, 443–448.
9
12900 J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010

Zinc Catalyst
A R T I C L E S
Figure 2. Model for the cometal-free ACA. LR and LS are either enantiomers or diastereomers.
Table 1. Diffusion Coefficients D (10-10 m2 s-1) and Volumes V
(Å3) of the Free Ligand and of the Zinc Complex Derived from a
1:1 Ratio (5:Et2Zn)a
entry
temperature [K]
Dligand
Vligand
Dcomplex
Vcomplex
nb
1
2
3
4
5
270
260
250
240
230
2.13
2.11
2.08
2.09
2.10
479
494
511
509
498
1.45
1.39
1.36
1.32
1.29
1506
1710
1856
1992
2136
2.8
3.1
3.4
3.7
3.9
a Measured as 0.02 M solutions in toluene-[D8]. b The multiple n is
based on the formula [ZnEtL]n.
successful formation of a zinc complex with a 1:1 stoichiometry.
At ratios higher than 1:1, the signal pattern of the zinc complex
does not change significantly, and additional signals of the free
uncomplexed diethylzinc appear at 1.13 and 0.06 ppm. It is
noteworthy that the line widths of the signals increase slightly.
This line broadening can be most probably attributed to an
exchange process of the ethyl substituent between the complex
and the uncomplexed diethylzinc, which is observed in NOESY
spectra (see SI). As a result, the NMR spectra indicate that
exclusively zinc complexes with the stoichiometry (ZnEtL)n are
formed in solution even in the presence of an excess of Et2Zn.
Diffusion-ordered 1H NMR experiments (DOSY) have been
successfully applied to obtain information about organometallic
aggregates and complexes in solution, providing a realistic
picture of microscopic dynamics.14 Therefore, DOSY experi-
ments were chosen to investigate the size of the complex
(ZnEtL)n, its aggregation level n and the temperature-dependence
of the aggregation. The diffusion coefficients of the free ligand
and of the complex were measured in a temperature zone
between 230 and 270 K, which lies in the synthetically relevant
area. For intensity reasons the concentration of the samples was
chosen to be double as high (0.02 M) as for the NLE studies
(0.01 M) to obtain reliable diffusion coefficients. The resulting
temperature- and viscosity-corrected diffusion coefficients are
summarized in Table 1.
Figure 3. Activity investigation of mixtures of diastereomeric catalysts
(Rp,S)-4 (corresponding to -100% de) and (Sp,S)-4 (corresponding to
+100% de). The calculated curve is based on parameters found by fitting
of the experimental values: see SI.
Figure 4. 1H NMR spectra of (Rp,S)-5 and diethylzinc in different ratios
(toluene-[D8] at 240 K).
First, 1H NMR spectra of the ligand and diethylzinc at
different ratios in toluene-[D8] were measured to identify the
stoichiometry of the complex (see Figure 4).
For the indicated temperatures, the diffusion coefficient of
(Rp,S)-5 (Dligand) remains constant within the experimental error
of (3%. The corresponding volumes indicate monomeric ligand
structures. For a monomeric zinc complex (ZnEtL) the diffusion
At a 1:1 ratio of (Rp,S)-5 and diethylzinc, the signals of the
free ligand completely disappeared, and a new set of signals
with different chemical shifts and broader line widths is
1
observed. The complete H and 13C chemical shift assignment
(14) (a) Henze, W.; Gärtner, T.; Gschwind, R. M. J. Am. Chem. Soc. 2008,
130, 13718–13726. (b) Schober, K.; Zhang, H.; Gschwind, R. M.
J. Am. Chem. Soc. 2008, 130, 12310–12317. (c) Zhang, H.; Gschwind,
R. M. Angew. Chem. 2006, 118, 6540–6544. Angew. Chem., Int. Ed.
2006, 45, 6391–6394.
of the zinc complex with the signals of the metal-bound ethyl
substituent (at 2.14 ppm (CH3) and 1.10 and 1.01 ppm
(diastereotopic CH2), see SI for further details) indicates the
9
J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010 12901

A R T I C L E S
Ay et al.
Scheme 2. (a) Possible Conformations of the Zinc Complex
ZnEt·5; (b) NOE Pattern of a 1:1 Complex of (Rp,S)-5 and
Diethylzinc Derived from a NOESY Spectrum at 260 Ka
coefficients would be similar to that of the free ligand, since
the additional volume of EtZn is relatively small. Surprisingly,
the diffusion coefficients of the 1:1 complex are significantly
smaller and show a slight temperature dependence, which
indicates aggregation of the complex in solution. On the basis
of the complex stoichiometry [ZnEtL]n (see above) and hard
sphere increments, an aggregation level of n ) 2.8 was
calculated for 270 K, which increases at lower temperatures
(up to n ) 3.9 at 230 K, see Table 1). In accordance with the
1
results of the H NMR spectra (Figure 4) also further DOSY
measurements show that increased amounts of Et2Zn up to 4:1
(Et2Zn to ligand) do not affect the structure or aggregation of
the complexes. The excess of Et2Zn causes only exchange peaks
between the free and bound EtZn in EXSY spectra, but the
diffusion coefficients of the resulting complexes are nearly
identical to those of the 1:1 samples (see SI). The presented
data show that despite the bulky [2.2]paracyclophane ligand the
zinc complex of 5 is aggregated and not monomeric in toluene.
The missing self-aggregation of the free ligand compared to
the [ZnEtL]n complex indicates a Zn-mediated aggregation
mechanism in solution. The nearly linear slope of the aggrega-
tion level n with decreasing temperature hints at an aggregation
structure with sticky ends.15 Another explanation would be a
very flat hyperpotential surface of the dimeric, trimeric, and
tetrameric aggregates for the complexes of (Rp,S)-5. In contrast,
for equilibria between strongly defined aggregates, plateau
values of the diffusion coefficients are often observed.16
a The color of the arrows corresponds to the intensity of the cross peaks
(red: very strong, blue: strong, green: weak).
As a result the DOSY investigations show, despite the bulky
paracyclophane ligand (Rp,S)-5, a distinct tendency for the
aggregation of the complex (ZnEt ·5)n in toluene with aggrega-
tion numbers between 2.8 and 3.9. To this increased aggregation
of (ZnEt ·5) compared to the NLE studies with (ZnEt ·4) two
spectroscopically necessary deviations in the experimental setup
may contribute. First, higher concentrations, as used in the NMR
compared to the NLE studies, are well-known to induce a higher
degree of aggregation in solution.17 Second, the aggregation
level of paracyclophane zinc complexes might be sensitive to
small changes in the ligand structure as similarly found in
phosphoramidite copper complexes.18 Thus, the lower sterical
hindrance and the reduced amount of entropic degrees of
freedom of the phenyl ring in 5 (necessary for the chemical
shift dispersion in the NMR studies) compared to the cyclohexyl
ring in 4 (used in the NLE studies) may additionally cause a
slightly higher aggregation in solution. Thus, in agreement with
each other, NMR and NLE studies show both aggregated
[LZnEt]n complexes in solution with the actual aggregation level
n dependent on concentration and ligand structure.
(ZnEtL) units. Other complexes can be excluded. In principle
four possible conformations of these monomeric ZnEtL units
with respect to the position of the zinc atom and the phenylethyl
side chain are likely to exist (Scheme 2a). The zinc atom looks
either downward and is pushed from the coordinating imine-
nitrogen into the paracyclophane half-space, giving conformer
ZnEt·(Rp,S)-5a, or it looks up. In this case, it is above the
paracyclophane scaffold leading to conformer ZnEt ·(Rp,S)-5b.
In the other two conformations the phenylethyl-moiety is flipped
away from the paracyclophane half-space (structures ZnEt ·(Rp,S)-
5c and ZnEt ·(Rp,S)-5d). An interconversion of the four con-
formations at the considered temperature range is improbable
because of the steric hindrance. This is also experimentally
confirmed by a series of temperature-dependent 1H NMR spectra
(270-220 K) of a complex of (Rp,S)-5 and diethylzinc at a 2:1
ratio (data not shown). In this temperature range significant line
broadening of any signal of (ZnEt ·5) is not observed, indicating
the absence of dynamic processes within the ligand. In general,
conformations ZnEt ·5b and ZnEt ·5d can be considered im-
probable due to energetical discrimination. Because of steric
reasons conformation ZnEt ·5c with the side chain flipped away
from the zinc core seems to be favored in comparison to
ZnEt·5a.19 All these considerations designate configuration
ZnEt·5c being the most probable structure in solution. However,
these were not underpinned by NMR experiments for paracy-
clophane zinc complexes with aggregation levels higher than
dimers. In a NOESY spectrum of a 1:1 complex of (Rp,S)-5
and diethylzinc at 260 K, the diethyl substituent shows the
The observation of one set of NMR signals at all temperatures
investigated shows that in solution the paracyclophane zinc
complexes exist as symmetrically aggregated species with
(15) For example a ladder structure that has been principally postulated
for methylzinc alkoxide clusters by DFT calculations: Steudel, R.;
Steudel, Y. J. Phys. Chem. A 2006, 110, 8912–8924.
(16) (a) Schober, K.; Zhang, H.; Gschwind, R. M. J. Am. Chem. Soc. 2008,
130, 12310–12317. (b) Zhang, H. Structure Elucidation of Copper-
Phosphoramidite Complexes in Solution. Ph.D. Thesis; University of
Regensburg: Regensberg, Germany, 2007.
(17) (a) Xie, X.; Auel, C.; Heinze, W.; Gschwind, R. M. J. Am. Chem.
Soc. 2003, 125, 1595–1601. (b) Henze, W.; Vyater, A.; Krause, N.;
Gschwind, R. M. J. Am. Chem. Soc. 2005, 127, 17335–17342. (c)
Zuccaccia, D.; Macchioni, A. Organometallics 2005, 24, 3476–3486.
(18) (a) Zhang, H.; Gschwind, R. M. Chem.sEur. J. 2007, 13, 6691–6700.
(b) Schober, K.; Hartmann, E.; Zhang, H.; Gschwind, R. M. Angew.
Chem. 2010, 122, 2855–2859. Angew. Chem., Int. Ed., 2010, 49,
2794-2797.
(19) Since the internal hydrogen bond of the free ligand can be exchanged
by the zinc moiety, it can be imagined that conformation 5c is favored
in comparison to 5a. The conformation 5c is favored by the free (Rp,S)-
ligand, which was shown by DFT-calculations recently: Warnke, I.;
Ay, S.; Bra¨se, S.; Furche, F. J. Phys. Chem. A 2009, 113, 6987–6993.
9
12902 J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010

Zinc Catalyst
A R T I C L E S
Scheme 3. Formation of the Complex Resulting from (Sp,R)-5 and
Dimethylzinca
a The NOE pattern hints for the shown conformation. The color of the
arrows corresponds to the intensity of the cross peaks (red: very strong,
blue: strong, green: weak).
strongest NOE cross peaks to one proton in the unsubstituted
paracyclophane ring, followed by the interaction with the
aromatic protons of the phenylethyl moiety (for the NOE
network see Scheme 2b). These observations verify clearly the
conformation ZnEt ·5a as most probable in solution. The other
weak NOE interactions are in agreement with this model (see
SI).20
Figure 5. Molecular structure of the methylzinc complex (ZnMe·(Sp,R)-
5)2. (Displacement parameters are drawn at 30% probability level; solvent
toluene and hydrogen atoms, except H19, H19′, are omitted for clarity.)
The reaction of Me2Zn with (Sp,R)-5 results in the formation
of the complex (ZnMe ·5), which was isolated and analyzed by
NMR (Scheme 3). A complex of a 1:1 stoichiometry was
identified as the sole product.21 Surprisingly, a slightly different
NOE pattern was observed in comparison to the diethylzinc
complex ZnEt · 5 (for details see SI). The NOESY spectrum
shows that the methyl substituent bound to zinc has a strong
cross peak to the unsubstituted paracyclophane ring, which
indicates that the zinc moiety is oriented downward. However,
the strongest cross peak can be assigned to the methyl substituent
(bound to zinc) and the methyl group of the phenylethyl side
chain. This is in contrast to the conformation of ZnEt ·5. This
interaction can only arise from the conformation ZnMe ·5c,
where the side chain is flipped away from the paracyclophane.
Fortunately, it was possible to isolate suitable crystals for
X-ray analysis of ZnMe ·(Sp,R)-5. The corresponding structure
is presented in Figure 5 and marks the first X-ray structure for
a zinc complex with a [2.2]paracyclophane ligand. It can be
reasoned that in the solid state a dimeric structure is present.
Because enantiomeric pure (Sp,R)-ligand was used for this
complex, only homodimers having a C2-symmetry can be
formed. The absolute configuration was confirmed crystallo-
graphically by refinement of Flack’s x-parameter (x )
0.030(11)).22 Both monomeric units are connected Via a Zn2O2-
ring system. Because of the bulky [2.2]paracyclophane ligands,
the connecting entity is distorted from planarity. Usually, methyl
zinc complexes with less bulky ligands possess planar ring
systems.23 Moreover, the methyl group directly connected to
zinc is pushed into the [2.2]paracyclophane plane. This verifies
conformation ZnMe ·(Sp,R)-5c as the favored one in the solid
state and in solution, in which the phenylethyl-moiety is flipped
away from this plane due to steric reasons.
The combination of NLE, NMR, and X-ray studies now
allows an insight into the aggregation and conformation trends
of paracyclophane zinc complexes. Despite the high steric
hindrance of the paracyclophane ligands, all studies consistently
show aggregation of the RZnL units, mediated by the zinc
moieties, and conformations with the zinc atoms pushed into
the paracyclophane half-space. However, the actual level of
aggregation and the conformation of the phenylethyl side chain
seem to be sensitive to the individual ligand structure and the
zinc reagents applied. For (Rp,S)-4 and (Sp,S)-4 in combination
with Et2Zn and for ZnMe ·(Sp,R)-5 dimeric structures were found
in solution and in the solid state, whereas ZnEt ·(Rp,S)-5 shows
aggregation numbers between 3 and 4. In agreement with the
deviating aggregation level of ZnEt ·(Rp,S)-5 and ZnMe ·(Sp,R)-
5, the ligands in the two complexes adopt the different
phenylethyl conformations 5a and 5c, respectively.
Quantum chemical calculations mainly based on density
functional theory (DFT) were performed to obtain the relative
energies of several dimers and monomers of ZnMe ·5 using the
program package TURBOMOLE.24 The details of the calcula-
tions are given in the paragraph “Quantum chemical calcula-
tions” in the SI. The results are summarized in Table 2;
enantiomers are listed only once. We compare DFT results
obtained with different exchange correlation functionals, and
consider the influence of van der Waals (vdW) effects by the
DFT-D method25 and solvent effects by the COSMO method.26
In all calculations for the monomers, the (Sp,R) monomer
(used in the experiment above) was about 4 kcal/mol more stable
than the (Sp,S) isomer. The relative energies of the dimers reflect
the energy differences of the monomers, the Sp,R/Sp,R dimer
has the lowest and the Sp,S/Sp,S dimer the highest energy. The
results of the DFT calculations are comparable for different
exchange correlation functionals. For the dimers, the energy
differences were increased, when van der Waals (vdW) interac-
tions were considered by the DFT-D method.27
(20) For a detailed discussions of X-Ray structures Vs NMR structures: (a)
John, M.; Auel, C.; Behrens, C.; Marsch, M.; Harms, K.; Bosold, F.;
Gschwind, R. M.; Rajamohanan, P. R.; Boche, G. Chem.sEur. J. 2000,
6, 3060–3068. (b) Xie, X.; Auel, C.; Henze, W.; Gschwind, R. M.
J. Am. Chem. Soc. 2003, 125, 1595–1601. (c) For copper-complexes:
Schober, K.; Zhang, H.; Gschwind, R. M. J. Am. Chem. Soc. 2008,
130, 12310–12317.
An important question was, whether the catalyst is more stable
in its monomeric or dimeric form. Therefore, the dimer binding
(21) The aggregation of this complex in solution was not investigated by
DOSY. This material was used for X-ray analysis.
(24) TURBOMOLE, version 6.0; TURBOMOLE GmbH: Karlsruhe, Ger-
many, 2008, www.turbomole.com.
(22) Flack, H. D. Acta Crystallogr. 1983, A39, 876–881.
(23) (a) Meyer, N.; Lo¨hnwitz, K.; Zulys, A.; Roesky, P. W.; Dochnahl,
M.; Blechert, S. Organometallics 2006, 25, 3730–3734. (b) Birch, S. J.;
Boss, S. R.; Cole, S. C.; Coles, M. P.; Haigh, R.; Hitchcock, P. B.;
Wheatley, A. E. H. Dalton Trans. 2004, 3568–3574.
(25) (a) Grimme, S. J. Comput. Chem. 2004, 25, 1463–1473. (b) Grimme,
S. J. Comput. Chem. 2006, 27, 1787–1799.
(26) Klamt, A.; Schuerman, G. J. Chem. Soc., Perkin Trans. 1993, 2, 799–
805.
(27) Ha¨ttig, C.; Weigend, F. J. Chem. Phys. 2000, 113, 5154–5161.
9
J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010 12903

A R T I C L E S
Ay et al.
Table 2. Relative Energies of Monomers and Dimers of ZnMe ·5,
Dimer Binding Energies (in kcal/mol) for Various Diastereomers of
ZnMe·5a
phenyl rings which can interact with the toluene molecules. This
interaction is not included in the COSMO model, but can be
estimated in the following way. According to the work of
Sinnokrot and Sherrill29 the π-π sandwich interaction is in the
order of -2.3 kcal/mol at an optimal distance of 3.8 Å. Both
dimers can at least interact with six toluene molecules, each
phenyl ring with one solvent molecule. In contrast to the
heterodimer, the homodimer possesses an extra pocket between
the two phenyl rings in the side chain. The distance between
them is around 7.5 Å so that there is a place for at least one
additional toluene ring. This extra stabilization has been
estimated on the basis of a small model system containing two
benzene rings and a toluene molecule in the middle. A binding
energy of -7.3 kcal/mol was obtained at DFT-D level (BP86,
def2-TZVP basis set).
relative energies of the different isomers (in kcal/mol)b
monomers
BP86
BP86/COSMO
B3-LYP
BP86/DFT-D
Sp,R
Sp,S
0.00
3.75
0.00
3.96
0.00
3.93
0.00
3.94
dimers
BP86
BP86/COSMO
B3-LYP
BP86/DFT-D
Sp,R/Sp,R
Sp,R/Sp,S
Sp,S/Sp,S
Sp,S/Rp,R
Sp,S/Rp,S
Sp,R/Rp,S
0.00
1.72
3.81
2.80
1.94
1.23
0.00
1.55
3.82
2.66
1.89
1.19
0.00
1.62
3.76
2.78
1.90
1.22
0.00
5.44
9.21
4.13
3.72
2.33
In summary, the calculations confirm the experimental finding
that the catalyst is mainly in its dimeric form. However, in
solution, the stabilization of the dimers will be less pronounced,
because of dispersion interactions of the monomer with solvent
molecules, which can fill the space occupied by the second
molecule in the dimer. The (Sp,R) homodimer has been found
to be the most stable dimer and is favored mainly by entropy
contributions and solvent interaction.
binding energies of the dimers (in kcal/mol)b
isomer
BP86
BP86/COSMO
B3-LYP
BP86/DFT-D
Sp,R/Sp,R
Sp,R/Sp,S
Sp,S/Sp,S
Sp,S/Rp,R
Sp,S/Rp,S
Sp,R/Rp,S
2.84
0.86
-0.93
-1.76
-0.39
-0.40
4.92
2.77
0.94
0.77
1.96
1.88
0.27
-1.76
-3.72
-5.32
-4.41
-4.83
-25.54
-25.94
-26.24
-25.94
-25.71
-25.57
a All values were obtained with a def2-TZVP basis set. All values
3. Conclusions
given include the ZPE corrections. The binding energies are corrected
b
In this study first indications were found to understand the
cometal-free asymmetric 1,4-selectivity of paracyclophane
ligands in the addition of dialkylzinc to enals. In accordance to
the well-established ML2-systems found for similar Zn-catalyzed
reactions using bidentate ligands such as DAIB, NMR studies
of paracyclophane zinc complexes including DOSY investiga-
tions and the first X-ray structure of a zinc complex with a
[2.2]paracyclophane ligand show consistently the existence of
(ZnL)n complexes. The observed negative NLE and the X-ray
analysis as well as the theoretical studies are in agreement with
the formation of dimeric species (ZnL)2 that serve as precatalysts.
In addition, the NLE investigations of mixtures of diastere-
omeric catalysts indicate the existence of additional possible
aggregates, which are very likely oligomeric (ZnL)n complexes
as found in the DOSY investigations of very similar complexes.
The presented results from NLE, NMR, and X-ray investigations
show that the level of aggregation n in these (ZnL)n complexes
is very sensitive to temperature and concentration.
for the basis set superposition error. Ebind ) EAB - EA - EB - zpeA
-
zpeB + zpeAB (zpe ) zero-point energy; EAB includes the counterpoise
correction).
energies calculated by different methods are summarized in the
last part of Table 2. As long as vdW interactions were not taken
into account, the energies of monomers and dimers were similar.
Inclusion of solvent effects (COSMO) stabilized all monomers
by 2 kcal/mol. By changing the functional to B3-LYP, the
dimers were slightly stabilized. When vdW interactions were
considered by the DFT-D method, the dimers were strongly
stabilized by about 25 kcal/mol. The strong influence of dynamic
correlation was confirmed by RI-MP2 calculations for the
binding energy of the (Sp,R) homodimer. For this isomer the
total binding energy amounts to -32.9 kcal/mol at MP2 level.
In the following, we focus our discussion on the homodimer
and heterodimer formed by the (Sp,R) and (Rp,S) monomers.
The calculated values show that the homodimers are around 2
kcal/mol more stable than the heterodimer. However, such a
stabilization is very small and prompted us to investigate which
factors affect the relative stability of the dimers. To gain insight
into the thermochemistry of the dimer formation the zero-point
energies were obtained from the frequency calculations and are
already included in the binding energies given in Table 2. The
As a proposal resulting from the presented investigations, the
1,4-selectivity of these complexes may be caused by the (ZnL)2
structure of the paracyclophane zinc complexes, which allows
cooperative clusters as found in the 1,4-addition reactions of
organocuprates or just enlarges the distance between coordinat-
ing Zn and the transferable alkyl moiety.
change in standard internal energy at 0 K (∆Uo(0) ) zpedimer
-
Σmonomer,i zpei) equals 1.3 kcal/mol for the homodimer and 6.1
kcal/mol for the heterodimer. At the reaction temperature of
248 K the homodimer has an entropic stabilization of 4.2 kcal/
mol compared to that of the heterodimer. While the reaction
proceeds in a nonpolar medium, the greater size of the polar
surface of one dimer will cause its destabilization in this solvent.
Topological molecular polar surface area28 (TPSA) calculations
give a value of 25 Å2 for the homodimer, while the value for
the heterodimer amounts to 33 Å2. A simple explanation of these
results is that the polar center of the homodimer is better
shielded. Moreover, the considered molecules contain several
4. Experimental Section
Crystal Structure Study of (ZnMe ·5)2. The single-crystal X-ray
diffraction study of (ZnMe ·5)2 was carried out on a Bruker-Nonius
APEXII diffractometer at 123(2) K using Mo KR radiation (λ )
0.71073 Å). Direct methods (SHELXS-97)30 were used for structure
solution, and refinement was carried out using SHELXL-97 (full-
matrix least-squares refinement on F2).30 H atoms were localized
by difference Fourier synthesis and refined using a riding model.
A semiempirical absorption correction was applied (max/min
(29) Sinnokrot, M. O.; Sherrill, C. D. J. Am. Chem. Soc. 2004, 126, 7690–
7697.
(28) Ertl, P.; Rohde, B.; Selzer, P. J. Med. Chem. 2000, 43, 3714–3717.
(30) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112–122.
9
12904 J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010

Zinc Catalyst
A R T I C L E S
transmission 0.8799/0.7515). The absolute configuration was
Union Road, Cambridge DB2 1EZ, U.K. (Fax: int.code +(1223)
336-033; E-mail: deposit@ccdc.cam.ac.uk).
confirmed by refinement of Flack’s x-parameter (x ) 0.030(11)).22
(ZnMe·5)2: yellow crystals, C54H58N2O2Zn2 - 0.5 toluene, M
) 943.83, crystal size 0.35 mm × 0.20 mm × 0.15 mm, trigonal,
space group P3121 (No. 152): a ) 13.2807(4) Å, c ) 52.5695(17)
Å, V ) 8029.8(4) Å3, Z ) 6, F(calcd) ) 1.171 Mg m-3, F(000) )
2982, µ ) 0.936 mm-1, 88125 reflections (2θmax ) 55°), 12234
unique [Rint ) 0.069], 599 parameters, 84 restraints, R1 (I > 2σ(I))
) 0.046, wR2 (all data) ) 0.120, GooF ) 1.09, largest diff. peak
Acknowledgment. This work has been supported by the Fonds
der Chemischen Industrie (fellowship for S.A.). K.F. was supported
by the Deutsche Forschungsgemeinschaft (DFG) through the Center
for Functional Nanostructures (CFN) in Karlsruhe.
Supporting Information Available: Time-dependence of the
product ee, mathematical models used for the fitted curves of
the NLE diagrams, NMR spectroscopy, NMR characterization
of the complexes and quantum chemical calculations; crystal-
lographic file in CIF format. This material is available free of
charge Via the Internet at http://pubs.acs.org.
and hole 0.509 and -0.390 e ·Å-3
.
Crystallographic data (excluding structure factors) for the
structure reported in this work have been deposited with the
Cambridge Crystallographic Data Centre as supplementary publica-
tion no. CCDC-723473 ((ZnMe ·5)2). Copies of the data can be
obtained free of charge on application to The Director, CCDC, 12
JA1032502
9
J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010 12905